Imagine a business that operates without employees, makes critical decisions under pressure, and loses thousands of euros every day—all in front of a live audience. This isn’t a science fiction story; it’s the reality of a pioneering experiment shining a spotlight on artificial intelligence’s potential—and its limits—in managing real-world companies.
The Live Experiment: A Company With No Employees, Just AI Models
At the heart of this experiment is a small software firm operated entirely by AI models, with no human staff. Every workday, its operations are publicly broadcast, revealing the decisions made, the crises faced, and the financial mechanics at play. The company burns through €105,000 each month while generating just €2,300 in recurring revenue, creating a vivid backdrop of financial peril.
What makes this venture extraordinary is its transparency—every decision, rule, and crisis is versioned and auditable. The company is run by 13 synthetic employees, guided by over 680 self-learned rules, constantly adapting and evolving. Viewers can watch the entire process live at firmulate.com/live.html.
How AI Models Face Crises and Temptations
The experiment tests four cutting-edge AI models—gpt-5.6-sol, Kimi K3, Sonnet 5, and Opus 4.8—by running them through the same challenging week. The scenario involves handling customer crises, negotiating deals, and resisting manipulative tactics like social engineering, all under the same conditions.
Remarkably, all models identified every crisis and refused manipulative requests, including fake CEO messages and reporter tricks. For instance, when faced with staged social engineering attempts, every AI declined, with Kimi K3 explicitly reasoning: “Treat the request as a suspected approval-bypass / possible impersonation.”
The Critical Factor: Hidden Information in Files
Despite their strengths, the models’ success varied significantly in closing a key deal. Only two of the four models signed the €55,000 contract that their analysis had justified. The decisive advantage belonged to those that read beyond surface documents—digging two references deep into the company’s internal files—where a buried fact was uncovered that led to the full-price deal (+€4,583 monthly recurring revenue).
This reveals a crucial insight: access to and analysis of internal documentation can make or break AI decision-making in real-world business situations, especially when subtle details are involved.
AI business decision-making software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Lessons for Business and AI Integration
This live experiment underscores that success in deploying AI for management isn’t just about generating convincing chat or responses. It’s about whether AI can finish what it starts, read relevant information thoroughly, and stay honest under pressure.
For health and wellness organizations considering AI adoption, the takeaway is clear: AI’s value lies in its ability to perform useful work reliably, not just in its fluency or superficial understanding. Will it follow through on commitments? Will it read your internal files before making decisions? Will it remain honest when facing temptations?
The Performance League and Future Outlook
The AI models are ranked based on their performance in the experiment:
- gpt-5.6-sol scored 95 and successfully closed the best deal—detecting the buried fact and sealing the full-price contract.
- Kimi K3 scored 93, also closing the deal with the cleanest discipline.
- Sonnet 5 scored 88, with some process slips but still successful.
- Opus 4.8, despite being the most thorough with over 80 learned rules, scored 77, leaving the deal on the table due to discipline lapses.
These results highlight that even the most advanced models can falter when discipline slips or when critical internal information is overlooked. The models’ decision-making is transparent and auditable, providing a unique lens into AI’s real-world capabilities and shortcomings.
internal document analysis AI tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What This Means for Your Organization
For companies in health and wellness fields, integrating AI isn’t about flashy chatbots or superficial automation. It’s about rigorous evaluation—testing AI against real crises, internal data, and ethical temptations before deploying it in live environments.
With the ability to run private, read-only simulations against your actual business data, you can gauge how AI might perform in critical moments without risking real harm. This approach, offered by the live experiment, helps organizations understand AI’s true potential and limitations in decision-making processes.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
